System and method for optical coherence tomography

The application relates to a system and to a corresponding method for optical coherence tomography.

Optical coherence tomography (OCT) is a method of measuring light-scattering specimens on their inside. Due to its light-scattering properties biological tissue is particularly suitable for diagnostic examination by means of OCT. Since for OCT relatively low light intensities are sufficient and the wavelengths of the light used mostly come within the near infrared range (750 nm to 1350 nm), unlike ionising X-ray diagnostics it does not contaminate biological tissue with radiation. It is therefore particularly significant for medicine and is roughly comparable to ultrasound diagnostics. With OCT, instead of sound, broadband light which has a very short coherence length is used. The running times of the light reflected on different boundary layers within the specimen are recorded with the aid of an interferometer. With OCT, typically resolutions higher by one to two orders of magnitude are to be achieved than with ultrasound, but the measuring depth achievable is considerably smaller. Due to optical scattering the cross-section images obtained only reach into the tissue up to a depth of a few millimeters. The currently most important areas of application of OCT are in ophthalmology, dermatology and the diagnosis of cancer. However, there are also some non-medical applications, such as e.g. in materials testing.

A generic system is known from W. Y. Oh et al., OPTICS EXPRESS Vol. 14, No. 19 (2006) 8675-8684 wherein the light emitted by a xenon arc lamp is injected into a Michelson interferometer and spectrally modulated there by shifting a reference mirror located on a piezoelectric converter. The light spectrally modulated in this way is conveyed to a Linnik interferometer, impinges on the specimen to be examined, and is reflected by the latter. The light reflected by the specimen is conveyed to a detector and collected by the latter. At this, the light impinging on the detector in the form of an interference pattern generally has a periodic shape with a plurality of periods.

For the most complete and most reliable possible recording of the time behaviour of the light striking the detector, the latter must be collected at a plurality of points in time. Therefore, reliable recording of the time behaviour of the interference pattern requires a relatively large amount of time. Since the maximum sampling rate of detectors is generally limited, interference patterns can moreover only be reliably sampled up to a specific frequency.

It is the object of the invention to specify a system and a corresponding method for optical coherence tomography wherein the times required for the most reliable possible recording of interference patterns are reduced.

This object is achieved by a system and a method for optical coherence tomography. The system includes: an interferometer which has a beam splitter and at least one reflector, the optical distance of which from the beam splitter can be changed at a speed, light being injected into the interferometer and light being emitted by the interferometer with which a specimen is irradiated, and a detector for collecting light which is reflected by the specimen, wherein the intensity of the light injected into the interferometer or the intensity of the light emitted by the interferometer is modulated with a modulation frequency f_Mwhich is not equal to the Doppler frequency f_D, the Doppler frequency f_Dbeing given by twice the ratio of the speed v of the change of the optical distance between the reflector and the beam splitter to the average wavelength λ₀of the injected light: f_M≠f_D=2·v/λ₀. The method includes: injected into an interferometer and emitted by an interferometer is light with which a specimen is irradiated, the interferometer having a beam splitter and at least one reflector the optical distance of which from the beam splitter is changed at a speed, collecting the light reflected by the specimen by a detector, modulating the intensity of the light injected into the interferometer or the intensity of the light emitted by the interferometer with a modulation frequency f_Mwhich is not equal to the Doppler frequency f_D, the Doppler frequency f_Dbeing given by twice the ratio of the speed v of the change of the optical distance between the reflector and the beam splitter to the average wavelength λ₀of the injected light: f_M≠f_D=2·v/λ₀.

The invention is based upon the idea of temporally modulating the intensity of the light injected into the interferometer or emitted by the interferometer, the modulation frequency of this modulation being different from the Doppler frequency. By means of the modulation of the intensity of the injected or emitted light, instead of a high-frequency interference pattern with a plurality of periods, a low-frequency beat frequency between the modulation and the interference pattern to be recorded is obtained on the detector, the low-frequency beat frequency having considerably fewer periods than the high-frequency interference pattern. Therefore, when recording this beat frequency with the detector, considerably fewer sampling time points per unit of time are required than when recording the interference pattern without the modulation of the light intensity according to the invention.

The modulation of the intensity of the light injected into the interferometer preferably takes place here before or during injection of the light, but can alternatively or additionally also take place after the injection in the interferometer.

The modulation of the intensity of the light emitted by the interferometer preferably takes place during the emission of the light on the outlet of the interferometer, but can alternatively or additionally also take place in the interferometer before the emission.

By means of the invention one gains the advantage that with a given maximum sampling rate of the detector considerably shorter times are required for reliable recording of an interference pattern obtained from a specific depth of the specimen. In this way the maximum number of interference patterns recordable with the detector per unit of time is considerably increased.

In comparison to a modulation with the Doppler frequency, by means of the modulation of the light according to the invention with a modulation frequency different from the Doppler frequency it is guaranteed that a low-frequency beat frequency is obtained independently of the respective phase position of the modulation of the intensity and of the time behaviour of the interference pattern. Due to the insensitivity of the system and method according to the invention to the respective phase position of the modulation and interference signal, acceleration of the recording of the interference pattern is guaranteed with high reliability.

The average wavelength of the light injected into the interferometer typically comes within the infrared spectral range, preferably between 750 and 1350 nm. In the case of a broadband light source the average wavelength of the light preferably comes within a spectral range in which the light source has an intensity maximum. Alternatively, the average wavelength is given by an average value from all of the wavelengths emitted by the light source.

The speed v is given by the temporal change of the optical distance between the reflector and the beam splitter. The optical distance between the reflector and the beam splitter is given by the spatial distance between the reflector and the beam splitter which is multiplied by the refraction index of the medium located between the reflector and the beam splitter. With an embodiment of the interferometer as a so-called free beam interferometer, with which air or a vacuum is located between the reflector and the beam splitter and the refraction index is approximately equal to 1, the optical distance of the reflector and the optical path by which the optical distance is changed is identical to the spatial distance or the spatial path of the latter. In this case the change of the optical distance of the reflector is implemented by a movement of the reflector by a spatial path. Alternatively, with an embodiment of the interferometer as a so-called fibre interferometer a light-conducting element, in particular an optical fibre, can be provided between the reflector and the beam splitter, the optical length of which can be changed specifically by an optical path. These optical fibres are also called fibre stretchers. In this case the optical distance and the optical path, by which the optical distance is changed, is given by the product of the spatial distance or the spatial path by which the distance is changed and the refraction index of the light-conducting element which typically comes within the range around 1.5.

Preferably, the modulation frequency with which the intensity of the injected or emitted light is modulated, is greater or smaller than the Doppler frequency by up to 40%: 1.2·v/λ₀≦f_M≦2.8·v/λ₀. By modulating the intensity of the light with a modulation frequency coming within this frequency range it is guaranteed.

It is preferred, moreover, that the detector has a maximum sampling rate which specifies the maximum possible number of sampling time points per unit of time when collecting the light striking the detector, and the maximum sampling rate corresponds to the frequency of the beat frequency signal or to a whole number multiple of the frequency of the beat frequency signal. In this way particularly fast and efficient sampling of the beat frequency signal is guaranteed.

A further embodiment of the design described here makes provision such that the frequency of the beat frequency signal, which is dependent upon the Doppler frequency and the modulation frequency, is set by specifying the Doppler frequency and/or the modulation frequency. Here, the Doppler frequency is preferably specified by specifying the speed of the change of the optical distance between the reflector and the first spectral splitter. In this way matching of the beat frequency signal to the maximum sampling rate of the detector, and so specification of the maximum number of interference patterns recordable per unit of time is made possible in an easy way.

In a further embodiment of the invention provision is made such that the modulation frequency corresponds to the maximum sampling rate of the detector or to a whole number multiple of the modulation frequency. In this way particularly fast and reliable recording of the interference pattern by the detector is guaranteed.

It is preferred that the maximum sampling rate of the detector is given by the reciprocal value of the minimum frame time of the detector which is composed of the minimum detection time which is required as a minimum when collecting the light impinging on the detector in the region of a sampling time point, and the minimum dead time which elapses until after collecting the light in the region of the sampling time point the light in the region of the next sampling time point can be collected. In connection with the preferred embodiments described above, with which the maximum sampling rate of the detector, the frequency of the beat frequency signal and the modulation frequency can be matched to one another in a way corresponding to the invention, with this embodiment matching of the frequency of the beat frequency signal and of the modulation frequency on the one hand and of the detector properties on the other hand is made possible.

Preferably, the intensity of the light which is injected into the interferometer or emitted by the interferometer is modulated sinusoidally or rectangularly. Both alternatives are particularly simple possibilities for intensity modulation.

In one preferred embodiment of the invention the system according to the invention comprises a light source for the emission of light which is injected into the first interferometer, and a modulation device, for example a chopper wheel, for modulating the intensity of the light emitted by the light source before the latter is injected into the interferometer. This embodiment is a particularly simple version for implementing a rectangular development of the intensity modulation.

Alternatively or additionally, a control device for controlling the light source is provided such that the intensity of the light emitted by the light source is modulated. In this way different developments of the intensity modulation, such as e.g. sinusoidal or rectangular developments, can easily be implemented.

In order to modulate the intensity of the light emitted by the interferometer an optical element is preferably provided which e.g. is disposed in the interferometer or in the region of the outlet of the interferometer, and can be specifically changed in its transmission or imaging properties. Thus, for example, by means of an adaptive optical element in the region of the outlet of the interferometer the intensity of the light emitted by the interferometer can be switched periodically from “high” to “low”. The optical element can, however, also be disposed in the optical path of the interferometer, e.g. between a reflector and the beam splitter.

Within the context of the invention, irradiation of the specimen with the light emitted by the interferometer is to be understood as meaning that the light emitted by the interferometer, which comprises the moveable reflector, impinges on the specimen directly or only impinges on the specimen after having passed through a further interferometer which is disposed between the interferometer and the specimen.

Within the context of the invention, collection of the light reflected by the specimen, in particular at different depths of the specimen, by the detector or the detector elements is to be understood as meaning that the detector or the detector elements collect the light from manifestations of interference which are produced upon superposition of the light reflected by the specimen, in particular at different depths of the specimen, with the light reflected on a reference mirror. The superposition of the light can take place here either in the interferometer which comprises the moveable reflector or in a further interferometer.

The invention and further advantageous embodiments of the invention are described in greater detail in the following by means of figures. These show as follows:

<figref idrefs="DRAWINGS">FIG. 1</figref> an exemplary embodiment of the OCT system according to the invention;

<figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>-<i>b</i>) two spatial elements of a specimen with individual sections;

<figref idrefs="DRAWINGS">FIG. 3</figref><i>a</i>-<i>b</i>) two cross-sections through the specimen and the specimen arm of the second interferometer;

<figref idrefs="DRAWINGS">FIG. 4</figref> a cross-section through the optical components of the second interferometer;

<figref idrefs="DRAWINGS">FIG. 5</figref> interference signals and the evaluation of the latter with the automatic calibration of the focus tracking;

<figref idrefs="DRAWINGS">FIG. 6</figref><i>a</i>-<i>c</i>) Interference signals and the envelope of the latter with non-modulated and modulated intensity of the light injected into the first interferometer;

<figref idrefs="DRAWINGS">FIG. 7</figref> an example of an electric circuit for modulating the sensitivity of the detector;

<figref idrefs="DRAWINGS">FIG. 8</figref> an exemplary structure of a so-called Linnik interferometer;

<figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>-<i>c</i>) three different positions of the specimen objective and the respectively obtained interference patterns;

<figref idrefs="DRAWINGS">FIG. 10</figref><i>a</i>-<i>b</i>) respective sections from a longitudinal section through the multimode fibre of the first optical fibre in the region of the input plane;

<figref idrefs="DRAWINGS">FIG. 11</figref> a section from a cross-section through the fibre bundle of the second optical fibre and a partial region of this section shown in enlarged form;

<figref idrefs="DRAWINGS">FIG. 12</figref> a section of the detector surface;

<figref idrefs="DRAWINGS">FIG. 13</figref> the detector surface and the inlet and outlet surface of the second optical fibre;

<figref idrefs="DRAWINGS">FIG. 14</figref><i>a</i>-<i>b</i>) two examples of the embodiment of the second optical fibre as cross-sections;

<figref idrefs="DRAWINGS">FIG. 15</figref> an interference pattern and a section from the interference pattern in comparison to the individual fibres of the second optical fibre;

<figref idrefs="DRAWINGS">FIG. 16</figref> a section from a longitudinal section through the fibre bundle of the second optical fibre in the region of the inlet surface;

<figref idrefs="DRAWINGS">FIG. 17</figref> a detector surface in the first operating mode;

<figref idrefs="DRAWINGS">FIG. 18</figref> a spatial element of the specimen with depth sections; and

<figref idrefs="DRAWINGS">FIG. 19</figref> a spatial element of the specimen with a two-dimensional tomogram at a specific depth.

<figref idrefs="DRAWINGS">FIG. 1</figref> shows an exemplary embodiment of the system according to the invention for OCT. The illustration chosen here of the individual components of the system is greatly schematised and not true to scale.

A first interferometer <b>10</b> has a first reference mirror <b>11</b> in a fixed position, a moveable second reference mirror <b>12</b> and a first beam splitter <b>13</b>. Light <b>14</b> from a light source <b>15</b> is injected into the first interferometer <b>10</b>, split by the first beam splitter <b>13</b> into a first partial beam <b>2</b> in the direction of the first reference mirror <b>11</b> in a fixed position and a second partial beam <b>3</b> in the direction of the moveable second reference mirror <b>12</b>. The two partial beams <b>2</b> and <b>3</b> are reflected by the fixed first reference mirror <b>11</b> and the moveable second reference mirror <b>12</b> and are superimposed in the first beam splitter <b>13</b> to form a third partial beam <b>4</b> which is injected into a first optical fibre <b>17</b> in the region of the output <b>8</b> of the first interferometer <b>10</b>, conveyed from the latter to a second interferometer <b>20</b> and injected here into an illumination arm <b>21</b> of the second interferometer <b>20</b>.

The light <b>14</b> injected into the first interferometer <b>10</b> is spectrally modulated by the optical path described in association with the movement of the second reference mirror <b>12</b> and leaves the first interferometer <b>10</b> in the form of the third partial beam <b>4</b> which is injected into the second interferometer <b>20</b>. Therefore, the first interferometer <b>10</b> can also be called a pre-modulator.

The second interferometer <b>20</b> serves as a sensor or measuring head which is brought manually by an operator, for example a doctor, into contact with the specimen <b>1</b> to be examined, in particular a biological tissue, and if appropriate is moved over the latter. The measuring head is so compact in structure here that its length preferably corresponds to that of a conventional writing implement such as e.g. a fountain pen.

In order to form the second interferometer <b>20</b> in this compact manner, the optical axes of the illumination arm <b>21</b> and of a reference arm <b>23</b> in which a third reference mirror <b>25</b> is in a fixed position, are respectively tilted about 90° in relation to the conventional perpendicular arrangement of the two optical axes (see the first interferometer <b>10</b>) and extend parallel to one another. In order to deflect the light beams from the illumination arm <b>21</b> and the reference arm <b>23</b> into the second beam splitter <b>24</b> a first and a second deflecting prism <b>26</b> and <b>28</b> are provided.

The first, second and third reference mirrors <b>11</b>, <b>12</b> and <b>25</b> do not have to be mirrors in the narrower sense, but are to be generally considered as surfaces which at least partially reflect the light located within the first and second interferometers <b>10</b> and <b>12</b>, and this is why the first, second and third reference mirrors <b>11</b>, <b>12</b> and <b>25</b> can also be called the first, second and third reflectors.

The partial beams superimposed in the second beam splitter <b>24</b> pass via the specimen arm <b>22</b> of the second interferometer <b>20</b> into the specimen <b>1</b>, are reflected here on boundary surfaces between media with different refraction indices, e.g. membranes or cell layers, and finally pass via the specimen arm <b>22</b> and the second beam splitter <b>24</b> into the output arm <b>27</b> from where they are injected into a second optical fibre <b>29</b> and conveyed via the latter to a detector objective <b>31</b> which images the light conveyed by the optical fibre <b>29</b> onto the surface of a two-dimensional detector <b>30</b>, enlarging it.

The detector <b>30</b> is preferably a semiconductor detector in CMOS technology and has a plurality of detector elements (pixels) disposed in an area, typically 640×512 pixels. Due to the simultaneous (“parallel”) recording of a plurality of reflections in different lateral positions from a plane at a specific depth of the specimen <b>1</b> made possible by this, this type of OCT can also be called “parallel OCT”.

The detector signals produced upon collecting the light striking the individual detector elements of the detector <b>30</b> are further processed in an electric circuit <b>32</b> and finally forwarded to a computer system <b>16</b> for graphic display and, if required, processing.

In comparison to OCT systems with just one interferometer, with the OCT system described here the movement of the second reference mirror <b>12</b> for the spectral modulation of the injected light <b>14</b>, the direct collecting of the light reflected by the specimen <b>1</b> and the recording of the image are allocated to three spatially separate components, namely to the first interferometer <b>10</b>, the second interferometer <b>20</b> which constitutes the measuring head, and the detector <b>30</b>.

By shifting the movement of the second reference mirror <b>12</b> and the recording of the image onto separate components, the second interferometer <b>20</b>, and so the measuring head, can be designed to be very compact and easy to manage. This makes the present OCT system particularly suitable for applications at external or internal locations of a body to be examined which are very difficult to access.

In the following sections preferred embodiments of the system according to the invention and advantageous combinations of individual embodiments are described in greater detail.

1. Depth Scan by Macroscopic Movement of the Reference Mirror

The moveable second reference mirror <b>12</b> in the first interferometer <b>10</b> has an optical distance I from the first beam splitter <b>13</b> and, starting from an initial position N, implements a linear, preferably periodic, movement towards the first beam splitter <b>13</b> and away from the first beam splitter <b>13</b> with an optical path length L and amplitude A, the optical path length L and the amplitude A being at least 100 times, preferably 1000 times, greater than the average wavelength λ₀of the light <b>14</b> injected into the first interferometer <b>10</b>.

The optical distance I is given here by the product of the spatial distance between the second reference mirror <b>12</b> and the first beam splitter <b>13</b> and the refraction index of the medium located between the second reference mirror <b>12</b> and the first beam splitter <b>13</b>.

With the preferred embodiment of the first interferometer <b>10</b> as a so-called free-beam interferometer described here with which air or a vacuum is to be found between the second reference mirror <b>12</b> and the first beam splitter <b>13</b> and the refraction index is approximately equal to 1, the optical distance I of the second reference mirror <b>12</b> and the optical path L by which the optical distance I is changed is identical to the spatial distance or spatial path of the latter. In this case the macroscopic change of the optical distance of the second reference mirror <b>12</b> is produced by a macroscopic movement of the second reference mirror <b>12</b> by a spatial path which is substantially greater than the average path length λ₀of the light <b>14</b> injected into the first interferometer.

Alternatively, with an embodiment of the first interferometer <b>10</b> as a so-called fibre interferometer (not shown) between the second reference mirror <b>12</b> and the first beam splitter <b>13</b> a light-conducting element, in particular an optical fibre, can be provided the optical length of which can be changed specifically by an optical path. These optical fibres are also called fibre stretchers. In this case the optical distance or the optical path by which the optical distance is changed is given by the product of the spatial distance or the spatial path by which the distance is changed and the refraction index of the light-conducting element which typically comes within the range around 1.5.

The average wavelength λ₀of the light <b>14</b> injected into the first interferometer <b>10</b> comes typically within the infrared spectral range, preferably between 750 and 1350 nm.

In the case of a broadband light source <b>15</b> the average wavelength λ₀of the light <b>14</b> preferably comes within a spectral range in which the light <b>14</b> of the light source <b>15</b> has an intensity maximum. Alternatively, the average wavelength λ₀is given by an average value of all of the wavelengths emitted by the light source <b>15</b>.

Preferably, the average wavelength λ₀of the light <b>14</b> injected into the first interferometer <b>10</b> comes within a wavelength range in which the detector <b>30</b> has a very high, in particular the highest, sensitivity. In the system illustrated, the light <b>14</b> has an average wavelength λ₀of approximately 1300 nm and a full width at half maximum (FWHM) of approximately 200 nm.

With an average wavelength λ₀of the light <b>14</b> in the range of e.g. 1 μm the optical wavelength L and amplitude A of the movement of the reference mirror <b>12</b> is therefore at least approximately 0.1 mm, preferably at least approximately 1 mm.

Unlike the normal microscopic amplitude of the reference mirror movement in the prior art in the order of magnitude of fractions of the average wavelength λ₀of the injected light <b>14</b>, i.e. of up to typically 1 μm, in the system described a macroscopic movement of the second reference mirror <b>12</b> in the order of magnitude of 0.1 mm to a number of millimeters is implemented.

During the macroscopic linear movement of the second reference mirror <b>12</b> the light reflected by the specimen <b>1</b> is forwarded via the second interferometer <b>20</b>, the second optical fibre <b>29</b> and the detector optics <b>31</b> to the two-dimensional detector <b>30</b> and recorded by the latter successively at a number of points in time respectively for a specific period of time which corresponds to the integration time of the detector <b>30</b>, and converted into corresponding detector signals.

In order for interference to be able to occur between the light reflected by the third reference mirror <b>25</b> and the light reflected by the specimen <b>1</b>, the so-called coherence condition must be fulfilled which, among other things proves that the respectively reflected light waves must have a constant phase relationship with one another in order to be able to interfere with one another. Due to the use of light <b>14</b> with a very short coherence length of typically 10 μm, the condition of a constant phase relationship is only fulfilled at specific depths or depth ranges of the specimen <b>1</b> which are therefore also called a coherence gate.

Each position of the second reference mirror <b>12</b> during the macroscopic movement corresponds here to a specific depth within the specimen <b>1</b> or a depth range around this specific depth for which the coherence condition is fulfilled so that interference can occur between the light reflected by the third reference mirror <b>25</b> and the light reflected by the specimen <b>1</b>.

In the case of a periodic movement of the second reference mirror <b>12</b> both half periods of the periodic movement of the second reference mirror <b>12</b> can respectively be used to record detector signals.

In this way successive two-dimensional sections are recorded from different depths of the specimen <b>1</b> by the detector <b>30</b>. This is illustrated in <figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>) in which—representative of a plurality of two-dimensional sections—a first, second and third two-dimensional section <b>34</b>, <b>35</b> and <b>36</b> are illustrated by a spatial element <b>33</b> of the specimen <b>1</b>. This type of two-dimensional section “passes” synchronously with the macroscopic movement of the second reference mirror <b>12</b> in direction a through the examined spatial element <b>33</b> of the specimen <b>1</b> without the latter having to be moved itself.

Each section <b>34</b>, <b>35</b> and <b>36</b> lies at a depth T<b>1</b>, T<b>2</b> and T<b>3</b> of the specimen <b>1</b> in which the coherence condition is respectively fulfilled so that interference can occur between the light reflected by the third reference mirror <b>25</b> and the light reflected by the specimen <b>1</b>. Therefore, the macroscopic movement of the second reference mirror <b>12</b> in combination with the successive two-dimensional collection of the light reflected by the specimen <b>1</b> has the effect of a three-dimensional depth scan.

<figref idrefs="DRAWINGS">FIG. 2</figref><i>b</i>) shows in comparison to this a method used in the prior art. In order to obtain sections <b>37</b> of different depths through the spatial element <b>33</b> observed the specimen <b>1</b> itself must be moved in direction b relative to the interferometer while the absolute position of the section <b>38</b> within the space remains substantially unchanged.

In contrast to this, the combination described above of the macroscopic linear movement of the reference mirror <b>12</b> on the one hand with the collecting of the light reflected by the specimen <b>1</b> with a two-dimensional detector <b>30</b> on the other hand enables recording of a complete three-dimensional set of data for the desired spatial element <b>33</b> of the specimen <b>1</b> which is substantially easier and quicker to implement. By means of the macroscopic movement of the second reference mirror <b>12</b> a three-dimensional tomogram is thus obtained instead of an image from a specific depth which is only two-dimensional. Unlike systems according to the prior art, with this method for recording a three-dimensional set of data the specimen <b>1</b> no longer needs to be moved relative to the second interferometer <b>20</b>. This makes the OCT system described compact, reliable and easy to handle, and so the latter is particularly suitable for use in vivo.

The three-dimensional set of data obtained in this way enables a precise diagnosis, in particular with biological specimens. Software-supported diagnosis aids can be used with a particularly high level of efficiency here, e.g. so-called “3D rendering” with which a three-dimensional set of data is processed by special software such that a quasi three-dimensional image is produced on a two-dimensional monitor. For this, cavities or tissue detachments, for example, can be shown as a three-dimensional animation—comparable to computer tomography (CT).

2. Focus Tracking

The OCT system described above is designed such that during a complete stroke, i.e. the wavelength L or twice the amplitude A, of the movement of the second reference mirror <b>12</b> an interference signal with sufficiently high intensity and great sharpness is always obtained. By means of the focus tracking described in greater detail in the following it is guaranteed that the interference signal and the sharpness of the interference pattern recorded are maximal for all depths within the specimen <b>1</b>.

For this purpose, while collecting the light reflected by the specimen <b>1</b> the focus, i.e. the focal point of the imaging optics on the side of the specimen of the second interferometer <b>20</b> is set such that the position of the focus in the specimen <b>1</b> and the position of the plane in the specimen <b>1</b> with which, in the case of reflection of light, the coherence condition is fulfilled and interference occurs, are substantially identical at all times while recording a tomogram of the spatial element <b>33</b> of the specimen <b>1</b>. This is illustrated in the following by means of <figref idrefs="DRAWINGS">FIGS. 3</figref><i>a</i>) and <b>3</b><i>b</i>).

<figref idrefs="DRAWINGS">FIG. 3</figref><i>a</i>) shows the case where the focus F of the specimen objective <b>41</b>—only shown in simplified form here as a lens—of the specimen arm <b>22</b> lies at a depth of the specimen <b>1</b> which does not correspond to the position of the coherence gate K. The section through the specimen <b>1</b> recorded within the coherence gate K at the depth T<b>1</b> is in this way not imaged exactly sharply onto the detector <b>30</b> (see <figref idrefs="DRAWINGS">FIG. 1</figref>), and so information losses during recording of the interference have to be accepted.

On the other hand, <figref idrefs="DRAWINGS">FIG. 3</figref><i>b</i>) shows the case where the focus F of the specimen objective <b>41</b> has been set such that it comes within the coherence gate K at the depth Ti. This tracking of the focus F of the specimen objective <b>41</b> corresponding to the respective depth Ti of the coherence gate K is called focus tracking. In this way, during the depth scan the second interferometer <b>20</b> is adjusted sharply to the respective position of the coherence gate K at different depths Ti of the specimen <b>1</b> so that images with great sharpness are obtained from each depth of the specimen <b>1</b>.

The maximum optical scan depth Tm specifies to which depth beneath the surface of the specimen <b>1</b> the coherence condition is fulfilled for constructive interference and corresponding interference patterns are obtained.

Moreover, by means of the focus tracking it is achieved that the illuminated surfaces on the unmoveable third reference mirror <b>25</b> in the second interferometer <b>20</b> on the one hand and at the respective depth of the specimen <b>1</b> on the other hand are identical at every depth Ti in the specimen <b>1</b> sampled. Moreover, the images of the respective illuminated surfaces via the reference arm <b>23</b> and the specimen arm <b>22</b> in the common image plane <b>27</b><i>a </i>of the reference and specimen arm <b>23</b> and <b>22</b> are identical and exactly superimposed.

In the following preferred embodiments of the OCT system described for the implementation of focus tracking are described in greater detail.

<figref idrefs="DRAWINGS">FIG. 4</figref> shows a cross-section through the arrangement of the individual optical components in the second interferometer <b>20</b>. The specimen objective <b>41</b> in the specimen arm <b>22</b> preferably comprises a number of lenses <b>42</b> which can be moved individually and/or in groups in direction R over the specimen <b>1</b> or away from the latter. For this purpose a piezoelectric actuator <b>40</b>, in particular an ultrasound piezo motor, is provided which is coupled to the specimen objective <b>41</b> or the lenses <b>42</b> and moves the latter along one or a number of guides <b>38</b>, in particular guide bars or guide grooves.

The movement of the lenses <b>42</b> preferably takes place synchronously with the macroscopic movement of the reference mirror <b>12</b> in the first interferometer <b>10</b> (see <figref idrefs="DRAWINGS">FIG. 1</figref>). In this way the focus F of the specimen objective <b>41</b> follows the coherence gate G while the latter passes through successive different depths T<b>1</b>, T<b>2</b> and T<b>3</b> of the specimen <b>1</b> from which, with the aid of the detector <b>30</b>, two-dimensional sections <b>34</b>, <b>35</b> and <b>36</b> (see <figref idrefs="DRAWINGS">FIG. 2</figref>) are respectively recorded.

The synchronisation of the macroscopic movement of the reference mirror <b>12</b> and the focus tracking on the one hand in combination with a two-dimensional detector <b>30</b> on the other hand guarantees particularly simple and rapid recording of a plurality of sharp, two-dimensional image sections at different depths of the specimen <b>1</b> and so the recording of a complete, three-dimensional set of image data with high image quality.

Since the first interferometer <b>10</b> and the optical imaging in the specimen arm <b>22</b> are continuously matched to one another, the interference signals recorded by the detector <b>30</b> for each depth in the specimen <b>1</b> are maximal so that a very high signal to noise ratio is produced. Moreover, in this way it is ensured that the lateral resolution for all depths in the specimen <b>1</b> is optimal because the focus F of the image always comes within the coherence gate K. In this way true-to-detail OCT images with high contrast are obtained.

Advantageously, the speed v<b>2</b> of the movement of the lenses <b>42</b> of the specimen objective <b>41</b> in direction R is lower than the speed v<b>1</b> of the movement of the reference mirror <b>12</b>. Preferably, a ratio v<b>1</b>/v<b>2</b> of the speeds of the reference mirror <b>12</b> and of the lenses <b>42</b> is chosen here which is approximately equal to 2·n−1 or up to approximately ±20%, preferably up to approximately ±10% around this value. In this way the position of the focus F and coherence gate G are matched to one another with a particularly high degree of reliability, as can be illustrated by the following consideration.

The focus F of the specimen objective <b>41</b> comes within a specimen <b>1</b> the refraction index n of which is generally not equal to one. If on the one hand one shifts the specimen objective <b>41</b> by a specific path in direction R of specimen <b>1</b>, the focus F shifts within the specimen by a specific amount d_F. For example, the shift of the specimen objective <b>41</b> by 0.78 mm with a refraction index of the specimen <b>1</b> of 1.4 leads to a shift in the focus in the specimen <b>1</b> by approximately d_F=1 mm. If, on the other hand, the reference mirror <b>12</b> is shifted by a specific path, the coherence gate K also shifts by a specific amount d_k. For example, a shift in the reference mirror <b>12</b> by 1.4 mm with a refraction index n=1.4 produces a shift in the coherence gate K by approximately d_k=1 mm. In this way with a shift in the reference mirror <b>12</b> and the specimen objective <b>41</b> respectively by the same path, with a depth scan the coherence gate K and the focus F would move apart from one another over a macroscopic depth range.

By means of the selection described above of the ratio v<b>1</b>/v<b>2</b> of the speeds of the reference mirror <b>12</b> and of the lenses <b>42</b> it is guaranteed that during the depth scan the coherence gate K and the focus F lie over one another in the whole depth range observed. In the above example of a specimen with a refraction index n=1.4 the ratio v<b>1</b>/v<b>2</b> of the speeds comes within the range of approximately (2·1.4−1)±20%, i.e. between approximately 1.44 and 2.16, and is preferably approximately 2·1.4−1=1.8.

The synchronisation of the movement of the reference mirror <b>12</b> and of the lenses <b>42</b> preferably takes place such that the reference mirror <b>12</b> and the lenses <b>42</b> pass at a specific point in time through two different, pre-defined spatial points at respectively constant, pre-defined and different speeds v<b>1</b> and v<b>2</b>.

After passing through the spatial points the recording of the actual OCT signals up to the pre-defined depth in the specimen <b>1</b> starts. With a periodic forwards and backwards movement of the reference mirror <b>12</b> OCT signals can be recorded here both during the forwards and during the backwards movement of the reference mirror <b>12</b>. The synchronisation of the reference mirror <b>12</b> and the lenses <b>42</b> takes place here in the same way and is re-set after each change in direction.

The measuring head in which the specimen objective <b>41</b> is located can be moved freely relative to the first interferometer <b>10</b> in which the second reference mirror <b>12</b> is located. A mechanical coupling of the specimen objective <b>41</b> and the reference mirror <b>12</b> for the synchronisation of the lens and reference mirror movements would lead to insufficient precision of the synchronisation.

Therefore, the synchronisation of the movements of the reference mirror <b>12</b> on the one hand and of the lenses <b>42</b> of the specimen objective <b>41</b> on the other hand is preferably implemented electronically. It is advantageous here to provide respectively in the region of the reference mirror <b>12</b> and the lenses <b>42</b> of the specimen objective <b>41</b> a position sensor <b>5</b> and <b>39</b> which records the current reference mirror and lens position and converts this into corresponding position signals. Both position signals are supplied to a control unit, in particular the computer system <b>16</b>, which then correspondingly controls the actuation of the reference mirror <b>12</b> and the lenses <b>42</b>.

The control of the reference mirror <b>12</b> and of the lenses <b>42</b> is preferably implemented by feedback of the position signals by means of a so-called master-slave system. With this type of master-slave system a measured position value in a first positioning unit is the basis for a desired value of the control circuit for a second positioning unit. In the present case the measured position of the first positioning unit of the reference mirror <b>12</b> is multiplied by a factor smaller than 1 and supplied to the second positioning unit of the lenses <b>42</b> as a new desired value. In this way the relative position error between the moveable reference mirror <b>12</b> and the lenses <b>42</b> is minimised, even with a relatively large absolute positioning error of the first positioning unit. In this way both components are coupled to one another electronically, like by means of a mechanical gearing, and so this can also be called electronic gearing.

The focus tracking can alternatively or additionally be implemented by an adaptive lens being provided in the specimen objective <b>41</b> the imaging properties of which can be specifically controlled and changed. For example, an oil and water lens can be controlled electrically such that the radii of curvature of the latter change, by means of which the focus of the latter can be changed and easily be adapted to the respective position of the coherence gate. In this case the speed and the start of the change of the focus F of the adaptive lens must be synchronised with the movement of the reference mirror <b>12</b> in the same way as the methods described above.

3. Automatic Calibration of the Focus Tracking

On the specimen side end of the specimen arm <b>22</b> of the second interferometer <b>20</b> in the form of a measuring head a material layer <b>43</b> (see <figref idrefs="DRAWINGS">FIG. 4</figref>) is provided which is preferably made of sapphire glass. The material layer <b>43</b> is coated on the inside <b>44</b> with an anti-reflex layer and is preferably uncoated on the outside <b>45</b> on the specimen side.

The OCT system can be operated in a diagnosis mode and in a calibrating mode. In the diagnosis mode, which corresponds to the normal measuring operation, the specimen side outside <b>45</b> of the material layer <b>43</b> is coated with a so-called index matching gel and brought into contact with the specimen <b>1</b> to be examined, three-dimensional images of which are recorded. In the calibrating mode the position of the focus F of the specimen objective <b>41</b> relative to the coherence gate K is determined, the outside <b>45</b> of the material layer <b>43</b>, which is preferably in air during the calibrating process, serves as a reference surface.

In the calibrating mode the amplitude of the OCT signal, which is caused by a reflection of the light due to the passage of the light from the material layer <b>43</b> into air, is measured for different positions of the specimen objective <b>41</b>, the following procedural steps, which are illustrated by means of <figref idrefs="DRAWINGS">FIGS. 4 and 5</figref>, being implemented: <ul><li id="ul0001-0001" num="0000"><ul><li id="ul0002-0001" num="0101">a) the group of lenses <b>42</b> is brought into an initial position by being moved as close as possible to the second beam splitter <b>24</b>;</li><li id="ul0002-0002" num="0102">b) the group of lenses <b>42</b> is left in this position;</li><li id="ul0002-0003" num="0103">c) during a macroscopic movement of the second reference mirror <b>12</b> the amplitude Ai of the maximum of the interference signal is determined;</li><li id="ul0002-0004" num="0104">d) the group of lenses <b>42</b> is moved away by a few micrometers, typically 5 to 20 μm, from the second beam splitter <b>24</b> and kept in this position;</li><li id="ul0002-0005" num="0105">e) steps c) to d) are repeated for a number of different positions P<b>1</b> to P<b>11</b> of the lenses <b>42</b>, for each position P<b>1</b> to P<b>11</b> of the lenses <b>42</b> an amplitude A<b>1</b> to A<b>11</b> of the maximum of the respective interference signal being obtained;</li><li id="ul0002-0006" num="0106">f) the position P<b>9</b> of the group of lenses <b>42</b> where the amplitude A<b>9</b> is at its greatest is established;</li><li id="ul0002-0007" num="0107">g) steps c) to f) are repeated close to position P<b>9</b> of this maximum with a smaller step width, typically 0.5 μm to 5 μm, the position P<b>9</b>′ of the group of lenses <b>42</b> where the amplitude A<b>9</b>′ is at its greatest being established;</li><li id="ul0002-0008" num="0108">h) from the reference mirror movement assigned to this position P<b>9</b>′ of the group of lenses <b>42</b> position Xm of the moveable reference mirror <b>12</b> where the interference signal is maximal is established.</li></ul></li></ul>

Alternatively, the calibration can also be implemented such that the specimen objective <b>41</b> moves to the second beam splitter <b>24</b> during the calibration.

If the group of lenses <b>42</b> is located in position P<b>9</b>′ and the reference mirror <b>12</b> in position Xm, the coherence gate and the focus position are identical. The established positions P<b>9</b>′ and Xm are set in the diagnosis mode as the initial position of the lens or lenses or of the reflector.

In this way any changes in the OCT system are automatically corrected without any additional hardware being required for this. Even if the material layer is contaminated or coated with index matching gel, the method described would work because then the passage of the light from dirt to air or gel to air would be used. The method is very fast and only lasts for a few seconds. It can therefore be implemented frequently, by means of which high system reliability is guaranteed.

In order to further increase the precision of the calibration method described, an additional element made of glass or plastic—a so-called target—can be applied to the material layer. The method described above is then implemented for two or more depths within the additional element. In this way, not only can an offset of the reference points for the movement of the reference mirror <b>12</b> and of the lenses <b>42</b> be corrected, but also any non-linearity. With the calibrating method described above a number of reference surfaces are then used, a number of position pairs being determined for which the focus position and the coherence gate are identical. In this way, not only can a constant relative position error between the two positioning units be corrected, but any errors in the relative linearity or the relative speed of the two units can be corrected. Such errors can be produced e.g. by ageing of the position sensors <b>5</b> and <b>39</b> when, for example, the position sensitivity of one of the two position sensors <b>5</b> and <b>39</b> changes.

In summary it can be established that the dynamic synchronisation of the focus position and the coherence gate in the diagnosis mode of the OCT system described leads to a plurality of advantages with regard to image quality and reliability. With additional, in particular regular, use of the calibrating mode described this synchronisation can be guaranteed over a long period of time.

4. Modulation of the Intensity of the Light Source

With the OCT system described the interference pattern produced is recorded with the detector <b>30</b>, a corresponding interference signal being produced. The sampling rate of the detector <b>30</b> for sampling the interference signal must be chosen here so that the temporal variation of the interference structure can be recorded with sufficient accuracy. This generally requires high sampling rates if high speeds are to be achieved for a depth scan.

Since the individual periods of an interference structure must generally respectively be sampled at a number of points in time, the maximum possible scan speed in the direction of the depth of the specimen <b>1</b> is dependent upon the maximum possible sampling rate of the detector <b>30</b>. When using rapid detector arrays with a high level of spatial resolution, i.e. a large number of detector elements per unit of length, the maximum sampling rate is typically in the range of approximately 1 kHz. With an average wavelength of the injected light <b>14</b> of for example 850 nm this leads to a maximum speed for the depth scan of approximately 0.1 mm/s if four points per period of the interference structure are recorded.

<figref idrefs="DRAWINGS">FIG. 6</figref><i>a</i>) shows the development over time of a typical interference signal which is sampled at a sampling rate of respectively four sampling time points P per period. In the figure four such points within a period of the interference signal are drawn in as an example.

In order to increase the speed of the depth scan, in the present OCT system the intensity of the light <b>14</b> injected into the first interferometer <b>10</b> is temporally modulated. This modulation takes place periodically, the frequency of the latter being greater or smaller by a specific amount, preferably by up to 40%, than the Doppler frequency f_Dwhich is given by the average wavelength λ₀of the injected light <b>14</b> and the speed v of the moveable reference mirror <b>12</b>: f_D=2v/λ₀. Typical frequencies of this modulation come within the range between 1 kHz and 25 kHz.

Alternatively or additionally, the intensity of the light of the third partial beam <b>4</b> emitted by the first interferometer <b>10</b> can also be modulated with the modulation frequency f_Min order to achieve the advantageous effect described above. The modulation is preferably implemented here during the injection of the light of the third partial beam <b>4</b> into the first optical fibre <b>17</b> at the output <b>8</b> of the first interferometer <b>10</b>. However, the intensity modulation can also take place in the second interferometer <b>10</b> before the light of the third partial beam <b>4</b> is emitted. In order to modulate the intensity of the light emitted by the second interferometer <b>10</b> an optical element is preferably provided which is disposed e.g. in the first interferometer <b>10</b> or in the region of the output <b>8</b> of the first interferometer <b>10</b> and can be specifically changed as regards its transmission or imaging properties. Therefore, for example, by means of an adaptive optical element in the region of the output <b>8</b> of the first interferometer <b>10</b> the intensity of the light of the third partial beam <b>4</b> emitted by the first interferometer <b>10</b> can be periodically switched from “high” to “low”. However, the optical element can also be disposed in the optical path of the first interferometer <b>10</b>, e.g. between one of the reference mirrors <b>11</b> or <b>12</b> and the first beam splitter <b>13</b>.

The definite choice of modulation frequency is made dependently upon the average wavelength λ₀of the injected light <b>14</b> of the light source <b>15</b>, the desired scan speed of the depth scan and the maximum sampling rate of the detector <b>30</b>.

Preferably the modulation frequency is chosen such that it corresponds to the maximum sampling rate of the detector <b>30</b> or a whole number multiple of the latter. The maximum sampling rate is given here by the reciprocal value of the minimum frame time of the detector <b>30</b>. The minimum frame time of the detector <b>30</b> is made up of the minimum time required in order to record a complete image and the minimum down time of the detector <b>30</b> which elapses until the next image can be recorded. The minimum frame time generally increases as the size of the image recorded increases.

The form of the modulation of the intensity of the light <b>14</b> is preferably sinusoidal or rectangular. The latter form can be produced e.g. simply by means of a rotating chopper wheel <b>18</b> (see <figref idrefs="DRAWINGS">FIG. 1</figref>). Other possibilities are acousto-optic or electro-optic modulators or liquid crystal modulators. A direct modulation of the light source <b>15</b> is also possible, the latter being controlled such that it emits the light <b>14</b> with temporally modulated intensity.

A corresponding effect can alternatively or additionally be achieved by an optical element, which is disposed e.g. before or after the first beam splitter <b>13</b> (see <figref idrefs="DRAWINGS">FIG. 1</figref>), being switched as to its transmission or imaging property. Therefore, for example, by correspondingly connecting an adaptive optical element the injection efficiency of the third partial beam <b>4</b> into the first optical fibre <b>17</b> could periodically be switched from “high” to “low”.

The described modulation of the intensity of the injected light <b>14</b> with a modulation frequency deviating, preferably slightly, from the Doppler frequency produces a low-frequency beat frequency between the modulation and the interference signal.

<figref idrefs="DRAWINGS">FIG. 6</figref><i>b</i>) shows the time behaviour of a beat frequency signal obtained upon the basis of the modulation described of the injected light <b>14</b> which—like the interference signal in the example of <figref idrefs="DRAWINGS">FIG. 6</figref><i>a</i>)—is sampled at a sampling rate of respectively four sampling time points P per period. Upon sampling the beat frequency signal, due to the lower frequency of the latter considerably fewer sampling time points P per unit of time are required than with the sampling of the interference signal in <figref idrefs="DRAWINGS">FIG. 6</figref><i>a</i>) so that with a fixed sampling rate given by the choice of detector <b>30</b>, considerably higher speeds can be achieved for the depth scan.

A further advantage of this method is described in greater detail in the following.

The integration time of the detector <b>30</b> corresponds to the period of time over which the detector <b>30</b> collects and thus integrates the light hitting the detector elements in the region of a time P. The detector <b>30</b> is preferably operated such that the integration time is only marginally shorter than the frame time. The frame time is chosen here such that it corresponds exactly to the duration of a period of the modulation or to a whole number multiple of the latter. The beat frequency signal shown in <figref idrefs="DRAWINGS">FIG. 6</figref><i>b</i>) was obtained by integration over the duration of two periods of the modulation.

If one were to increase the scan speed without modulating the intensity of the light <b>14</b> as described above, the frame time—and so the integration time—of the detector <b>30</b> would have to become shorter because the Doppler frequency would increase and in this way sampling time points P lying closer together in time would be necessary. However, a shorter integration time would lead to a reduction of the photons collected per integration and per detector element, and this would lead to a reduction in the signal/noise ratio due to the so-called Schott noise resulting from the statistical nature of the photons. In order to improve the signal/noise ratio again, the intensity of the injected light <b>14</b> would have to be increased in proportion to the scan speed.

If, on the other hand, one increases the scan speed with the aid of the modulation of the intensity of the light <b>14</b> described above, the integration time can remain constant. There is only a light loss of 50% due to the modulation of the light <b>14</b>. With the preferred modulation frequency, which corresponds to twice the reciprocal value of a frame time, there is an increase by factor 8 of the speed. In this case four times less light intensity is required in order to achieve this increase in speed than in the case without modulation. The effects of the light loss amounting to 50% due to the modulation are in this way over-compensated.

With the described method, the required intensity of the light <b>14</b> of the light source <b>15</b> must—unlike direct sampling without beat frequency—therefore not be increased with the scan speed because in this case the integration time of the detector <b>30</b> can remain constant.

A further advantage of the light modulation is the reduction of the quantity of data for a complete three-dimensional depth scan. With the recording of a three-dimensional set of data with a lateral size of 512×640 pixels and a scan depth of 1 mm in a tissue with the refraction index n=1.4, approx. 6 Gbytes of data are produced. With the modulation of the intensity of the light <b>14</b> described above the quantity of data is reduced to 750 Mbytes.

Moreover, the directly obtained data must additionally be processed in order to display the image result. Here too the reduced quantity of data is very advantageous because in this way the processing time is considerably reduced, and so the image result is available more quickly.

Preferably the Doppler frequency and/or the modulation frequency are chosen such that a period of the resulting beat frequency signal is a whole number multiple of the minimum frame time of the detector <b>30</b>, i.e. such that the maximum sampling rate of the detector <b>30</b> is a whole number multiple of the frequency of the beat frequency signal.

If one chooses a period length of the modulation of the light <b>14</b> as a minimum frame time of the detector <b>30</b>, the scan speed increases by factor 4 in relation to the scan speed with non-modulated light <b>14</b>. If, however, one chooses a minimum frame time of two periods of the modulation, the scan speed increases by the factor 8.

<figref idrefs="DRAWINGS">FIG. 6</figref><i>c</i>) shows the envelope Eu and Em of the interference signal or beat frequency signal shown in <figref idrefs="DRAWINGS">FIGS. 6</figref><i>a</i>) and <b>6</b><i>b</i>) with unmodulated and modulated light <b>14</b>. Each point P′ of the envelope Eu and Em corresponds here to a sampling time P of the associated interference signal or beat frequency signal.

Information is deduced from the respective envelope Eu and Em from which initially one-, two- and finally three-dimensional images of the specimen <b>1</b> are put together. As trials have shown, by means of the intensity modulation implemented, despite the considerably smaller number of measuring points P and P′, no relevant information losses in comparison to a conventional system without intensity modulation occur.

Overall, by means of the described modulation of the intensity of the injected light <b>14</b> the maximum possible speed of the depth scan is multiplied without any significant information losses occurring when the signal is evaluated.

5. Modulation of the Sensitivity of the Detector System

The principle of the modulation of the intensity of the light <b>14</b> injected into the first interferometer <b>10</b> and of the light of the third partial beam <b>4</b> emitted by the first interferometer described above can be analogously applied to the sensitivity of the detector system which, among other things, comprises the detector <b>30</b> and the detector objective <b>31</b> by the sensitivity of the detector system, in particular of the detector <b>30</b>, being modulated for the light to be collected with a frequency which is preferably greater or smaller than the Doppler frequency f_Dby a specific amount, in particular by up to 40%.

The light reflected by the specimen <b>1</b> and striking the detector <b>30</b> is superimposed here with the modulated sensitivity of the detector system <b>30</b>, <b>31</b> so that when recording the interference pattern striking the detector <b>30</b> the detector <b>30</b> produces, instead of a high-frequency interference signal with a plurality of periods, a low-frequency beat frequency signal which has considerably fewer periods than the high-frequency interference signal. With the sampling of this beat frequency considerably fewer sampling time points are therefore required per unit of time than with sampling of the high-frequency interference signal without the modulation of the sensitivity of the detector system <b>30</b>, <b>31</b>.

The sensitivity of the detector <b>30</b> can be modulated e.g. directly or with a controllable electronic shutter disposed in front of the detector <b>30</b>. Alternatively or additionally, properties of an optical element in the detector system, such as e.g. the permeability of the detector objective <b>31</b>, can be modulated for the light reflected by the specimen <b>1</b>.

The mode of operation of the direct modulation of the sensitivity of the detector <b>30</b> is illustrated in greater detail by means of <figref idrefs="DRAWINGS">FIG. 7</figref> which shows a greatly schematised electric circuit. Each of the detector elements <b>80</b> of a CMOS detector can be illustrated in simplified form in the equivalent circuit diagram as a photodiode <b>81</b> which is pre-stressed with a voltage U<b>1</b>. An ohm resistor and a capacitor are optionally connected parallel to the photodiode <b>81</b>. By irradiating the detector element <b>80</b> with light, charge carriers are produced in the photodiode <b>81</b> which trigger a flow of current I<b>1</b> which is added up in an accumulator <b>82</b> of an electronic integrator <b>83</b>. By means of periodic switching on and off of this integration by means of a switch <b>84</b> which is controlled with the modulation frequency f_M, the amount of charge, and so the respectively currently recorded light intensity is modulated with the modulation frequency f_M. By means of a sample-and-hold step <b>87</b> the corresponding detector signal is picked up and delivered for further processing. The further switches <b>85</b> and <b>86</b> serve to control the restoration of the integration and the picking up of the detector signal.

In the same way as the modulation of the intensity of the injected or emitted light <b>14</b> and <b>4</b> described above, with this version too, instead of a high-frequency interference signal, a low-frequency beat frequency signal (see <figref idrefs="DRAWINGS">FIGS. 6</figref><i>a</i>) and <i>b</i>)) is obtained which can be sampled with considerably fewer sampling time points P without losing any relevant information here. With a given maximum sampling rate of the detector <b>30</b>, the consequence of this is that the maximum speed for a depth scan of the system can be increased by a multiple.

As with the modulation of the injected or emitted light <b>14</b> and <b>4</b> (see section 4), here too, by means of an appropriate choice of frequency of the modulation of the sensitivity of the detector system <b>30</b>, <b>31</b>, the scan speed is increased by factor 4 or even 8 in comparison with systems with constant detector sensitivity.

The speed of the movement of the second reference mirror <b>12</b> is in a fixed relationship to the frequency of the modulation of the sensitivity of the detector <b>30</b> and is preferably chosen such that over a period duration of the beat frequency signal produced a whole number amount of sampling time points, preferably four sampling time points, pass (see <figref idrefs="DRAWINGS">FIG. 6</figref><i>b</i>)).

The beat frequency signals sampled in this way must be processed again before a visualisation because the interference information is also contained in these signals. The essential information which is to be visualised is the amplitude and the depth position of the respective interference, not however the interference structure itself. For this purpose the beat frequency signal must be demodulated, i.e. the envelope of the beat frequency signal is determined (see Em in <figref idrefs="DRAWINGS">FIG. 6</figref><i>c</i>)).

Since the phase of the beat frequency signal is generally unknown and this can also be different for different beat frequency signals from different depths, a digital demodulation algorithm is used which is independent of the phase. Preferably, for the sampling of the interference signal with four sampling time points per period so-called 90° phase shift algorithms are used. In this way fast demodulation of the beat frequency signal is achieved.

6. Measuring Head with Asymmetrical Linnik Interferometer

In the following the structure of the measuring head, which comprises the second interferometer <b>20</b>, is illustrated in greater detail by means of <figref idrefs="DRAWINGS">FIGS. 4</figref>, <b>8</b> and <b>9</b>.

The second interferometer <b>20</b> is a so-called Linnik interferometer. <figref idrefs="DRAWINGS">FIG. 8</figref> shows an example of a typical structure of this type of Linnik interferometer with a beam splitter <b>77</b>, reference mirror <b>78</b>, detector <b>79</b> and specimen <b>70</b>. With this type of Linnik interferometer limits are basically set for miniaturisation, and this applies in particular to the diameters of the optical elements used such as e.g. the objectives <b>75</b> and <b>76</b> and the lenses <b>71</b> and <b>74</b>, and the geometric structure. The structure of the specimen and reference objective <b>75</b> and <b>76</b> and the distance q between the latter and the beam splitter <b>77</b> are substantially equal.

With the Linnik interferometer used in the present OCT system the distances between the specimen and reference objective <b>41</b> and <b>46</b> and the second beam splitter <b>24</b> (see <figref idrefs="DRAWINGS">FIG. 4</figref>) are generally not equal for all scan depths due to the focus tracking. In this way large relative optical path length differences (OPD) can occur between the image centre and the image edge of the specimen and reference image. The consequence of these can be that the spatial frequency of the interference structure to be recorded becomes greater than the resolution of the two-dimensional detector <b>30</b> due to which the interference can no longer be demonstrated, or only be demonstrated insufficiently reliably.

In order to avoid these disadvantages, in the second interferometer <b>20</b> of the present OCT system the specimen and reference objective <b>41</b> and <b>46</b> are designed differently (“asymmetrically”) and matched to one another, as illustrated in greater detail below by means of <figref idrefs="DRAWINGS">FIG. 4</figref>.

The distance p between the specimen objective <b>41</b>, in particular of the lenses <b>42</b>, and the second beam splitter <b>24</b> is chosen to be very small. For the upper scan position in which the light reflected by a section lying close to the surface of the specimen <b>1</b> (see <figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>) is collected, the distance p is preferably between 1 and 3 mm. In this way the diameters of the lenses <b>42</b> and <b>49</b> in the specimen and reference arm <b>22</b> and <b>23</b> are chosen to be very small with at the same time a large light yield.

A further group of lenses <b>47</b> in the output arm <b>27</b> forms together with the specimen and reference objective <b>41</b> and <b>46</b> the specimen and reference optics. The specimen and reference optics are telecentric on the side of the specimen <b>1</b> and of the third reference mirror <b>25</b>. Telecentric optics are characterised in that the object distance can be varied and the image size nevertheless remains constant. This is achieved by an aperture stop.

The numerical aperture for the imaging of the specimen <b>1</b> is relatively large, preferably approximately 0.3. However, the numerical aperture of the illumination of the specimen <b>1</b> is smaller than the numerical aperture for the imaging of the specimen <b>1</b>, and preferably has a value of 0.2. In this way, together with the telecentric design of the specimen and reference optics one gains the advantage of the light reflected on inclined specimen structures also being picked up by the specimen objective <b>41</b> because the acceptance angle of the specimen objective <b>41</b> is greater than the divergence angle of the illumination cone. If the numerical aperture for illumination and imaging were of equal size, however, with the reflection on inclined specimen structures less light would be picked up than with the reflection on structures which are perpendicular to the optical axis.

In the specimen arm <b>22</b> the smaller numerical aperture for the illumination is provided by the choice of illumination objective <b>48</b> in the illumination arm <b>21</b>. The numerical aperture in the reference arm <b>23</b> is equal to or somewhat larger than the numerical aperture of the illumination arm <b>21</b>. This is particularly advantageous with the folded Linnik interferometer used here because in this way the reference objective <b>46</b> can be adapted relatively easily to the specimen objective <b>41</b> and moreover can be produced compactly.

The optical path through the lenses <b>49</b> of the reference objective <b>46</b> (including any air spaces between the lenses <b>49</b>) is shorter than the optical path through the group of lenses <b>42</b> of the specimen objective <b>41</b>.

By means of these measures it is possible for the image field curvatures of the specimen arm and the reference arm <b>22</b> and <b>23</b> in the centre of the used scan depth to be largely identical. Moreover, it is guaranteed that the maximum optical path length difference (OPD) between the image centre and image edge on the upper and lower end of the depth scan is small enough in order to guarantee a spatial frequency of the interference structure which is small enough in order to fulfil the Nyquist condition with regard to the detector <b>30</b>. In this way the spatial frequency of the interference structures from different depths in the observed spatial element <b>33</b> of the specimen <b>1</b> is always smaller than the resolution of the two-dimensional detector <b>30</b>. The interference structures are in this way always recorded with a high degree of reliability at every depth of the observed spatial element <b>33</b> of the specimen <b>1</b>.

This is illustrated in <figref idrefs="DRAWINGS">FIGS. 9</figref><i>a</i>) to <i>c</i>) in which a specimen side section of the cross-section of the second interferometer <b>20</b> is shown at three different times during a depth scan.

At a first time (see <figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>)) the coherence gate K is in an upper layer <b>34</b> of the observed spatial element <b>33</b> of the specimen <b>1</b> (see <figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>)). Here the specimen objective <b>41</b> is a small distance away from the second beam splitter <b>24</b> and a relatively large distance away from the material layer <b>43</b> or the specimen <b>1</b>. The interference structure obtained here is shown in the right-hand part of <figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>) and has a period length which corresponds to the distance between two respective consecutive light or dark rings. This period length is greater than the centre-centre distance (pitch) of the individual detector elements (pixels) of the detector <b>30</b>, i.e. the spatial frequency of the interference structure, which corresponds to the reciprocal period length, is smaller than the resolution of the detector <b>30</b> which corresponds to the reciprocal centre-centre distance of the pixels of the detector <b>30</b>, by means of which the so-called Nyquist condition is fulfilled. In this way it is guaranteed that the interference structure can be reliably recorded by the detector <b>30</b>.

At a second time (see <figref idrefs="DRAWINGS">FIG. 9</figref><i>b</i>)) the coherence gate K is in a central layer <b>35</b> of the observed spatial element <b>33</b> of the specimen <b>1</b> (see <figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>)). The specimen objective <b>41</b> is in a position which is slightly further away from the second beam splitter <b>24</b> and somewhat closer to the material layer <b>43</b> than in <figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>)). In this case the interference structure has a greater period length than in <figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>)) so that at this time too the Nyquist condition is fulfilled.

At a third time (see <figref idrefs="DRAWINGS">FIG. 9</figref><i>c</i>)) the coherence gate K is in the deepest layer <b>36</b> of the observed spatial element <b>33</b> of the specimen <b>1</b> (see <figref idrefs="DRAWINGS">FIG. 2</figref><i>a</i>)). The specimen objective <b>41</b> is in a position which is even further away from the second beam splitter <b>24</b> and even closer to the material layer <b>43</b> than in <figref idrefs="DRAWINGS">FIG. 9</figref><i>b</i>). In this case the interference structure has approximately the same period length as at the time illustrated in <figref idrefs="DRAWINGS">FIG. 9</figref><i>a</i>) so that in this depth scan position too the Nyquist condition is fulfilled.

Due to the described asymmetrical embodiment of the specimen and reference objective <b>41</b> and <b>46</b>, different distances and optical paths p and r between the specimen and reference objective <b>41</b> and <b>46</b> and the second beam splitter <b>24</b> can be produced. In the example shown, in this way the specimen objective <b>41</b> at distance p can be brought close to the second beam splitter <b>24</b>, by means of which small diameters of the lenses <b>42</b> with a high light yield can be produced. At the same time the reference objective <b>46</b> can be disposed a considerably greater distance away r (r&gt;p) from the second beam splitter <b>24</b>, by means of which folding of the second interferometer <b>20</b> is made possible with which the reference and illumination arm <b>23</b> and <b>21</b> are respectively tilted about 90° in relation to their position in a non-folded Linnik interferometer (see <figref idrefs="DRAWINGS">FIG. 8</figref>) and in this way extend parallel to the specimen arm <b>22</b>.

In this way a very slim form of the measuring head is produced and at the same time it is guaranteed that the image on the detector <b>30</b>, which is produced by the reference and specimen optics, is of equal size and well superimposed for all scan depths.

By means of the embodiment of the reference objective <b>46</b> described above, part of the optical path which is required for folding is compensated. Therefore, the reference objective <b>46</b> is optically shorter than the specimen objective <b>41</b>. In this way the embodiment of the first interferometer <b>10</b> is simpler because in this way the two interferometer arms of the first interferometer <b>10</b> do not have to differ from one another so greatly in order to fulfil the coherence condition for the occurrence of interference.

The difference between the optical path lengths in the reference and the specimen arm <b>23</b> and <b>22</b> is preferably at least twice as great as the maximum scan depth Tm (see <figref idrefs="DRAWINGS">FIGS. 3</figref><i>a</i>) and <i>b</i>)). The maximum optical scan depth Tm specifies up to which depth beneath the surface of the specimen <b>1</b> the coherence condition for the occurrence of interference is fulfilled and corresponding interference patterns are obtained. In this way a clear and simple assignment of the position of the reference mirror <b>12</b> in the first interferometer <b>10</b> at a specific depth in the specimen <b>1</b> is guaranteed.

7. Single Mode Pre-Modulation and Multimode Fibre

With the embodiment of the first interferometer <b>10</b> preferred here in the so-called free beam optics, when using the conventionally used spatially short or incoherent light sources a relatively complex objective is required in the region of the output <b>8</b> of the first interferometer <b>10</b> in order to inject the outgoing light as efficiently as possible into the first optical fibre <b>17</b> and thus avoid light losses. In this way not only the optical structure of the second interferometer <b>20</b>, which is to be designed as compactly as possible for endoscopic applications, but also the structure of the optics of the first interferometer <b>10</b> are restricted. Moreover, if applicable, any required increase in the light output is restricted with the conventionally used spatially short or incoherent light sources.

In order to avoid these disadvantages, in the present OCT system one or a number of single mode light sources respectively having high spatial coherence, such as e.g. superluminescence diodes (SLEDs), short pulse lasers or supercontinuum lasers, are used as a light source <b>15</b>. The light <b>14</b> of the light source <b>15</b> is injected into the first interferometer <b>10</b>, only the so-called Gauss mode, which corresponds to a single mode, being transmitted. Only after passing through the first interferometer <b>10</b> is the spatial coherence of the injected light <b>14</b> destroyed by the light at the output <b>8</b> of the first interferometer <b>10</b> being injected into the first optical fibre <b>17</b> which has a very long multimode fibre.

A multimode fibre is a fibre the numerical aperture and characteristic diameter of which allows not just one fibre mode to be formed with a specific wavelength of the light, but makes it possible for many different fibre modes to be stimulated. Whether a fibre is a single mode fibre or a multimode fibre can be estimated using the so-called V figure V:

<maths id="MATH-US-00001" num="00001"><math overflow="scroll"><mrow><mi>V</mi><mo>=</mo><mrow><mfrac><mi>π</mi><mi>λ</mi></mfrac><mo>·</mo><mi>d</mi><mo>·</mo><mi>NA</mi></mrow></mrow></math></maths>

λ specifying the wavelength of the light injected into the fibre, d the characteristic diameter of the fibre and NA the numerical aperture of the fibre. The wavelength λ of the light injected into the fibre is preferably identical here to the average wavelength λ₀of the light <b>14</b> injected into the first interferometer. If the V figure is greater than approximately 2.4, this is a multimode fibre.

The multimode fibre preferably used in the first optical fibre <b>17</b> has typical lengths in the order of magnitude of approximately 100 m and is preferably predominantly wound onto a coil <b>19</b>, as indicated in <figref idrefs="DRAWINGS">FIG. 1</figref>. The characteristic diameter of the multimode fibre is preferably between approximately 200 μm and approximately 400 μm.

The very long, thin and preferably wound up multimode fibre can optionally be combined in the first optical fibre <b>17</b> with a relatively short, thick fibre (not shown) the diameter of which comes within the range of approximately one millimeter and the length of which comes within the range of meters.

By means of the destruction of the spatial coherence of the light of the single mode light source <b>15</b>, the light reflected by two different points in the specimen <b>1</b> is prevented from being able to interfere, and this is also referred to as so-called coherent cross-talk.

Moreover, efficient suppression of the coherent cross-talk leads to effective suppression of undesired scattered light which, in the case of a light source with high spatial coherence, would also contribute to interference, and consequently would lead to a blurred, washy image—similar to an image behind a pane of frosted glass. In the way described above efficient destruction of the spatial coherence is implemented, by means of which the detection of scattered light is greatly reduced and finally a sharp image is obtained.

The pre-modulation information produced in the first interferometer <b>10</b>, i.e. the spectral modulation of the injected light <b>14</b> brought about by the movement of the second reference mirror <b>12</b>, is not changed, however, when the light is transmitted by means of the very long multimode fibre of the first optical fibre <b>17</b>. This is guaranteed in that in the multimode fibre both arms of the first interferometer <b>10</b> produce identical modes with identical mode distribution and identical phases.

Each mode itself then transmits the pre-modulation information, the individual modes not being coupled with one another. This is achieved by the first and second partial beams <b>2</b> and <b>3</b> in the first interferometer <b>10</b> (see <figref idrefs="DRAWINGS">FIG. 1</figref>) being superimposed co-linearly and exactly in relation to a third partial beam <b>4</b> before they enter into the multimode fibre of the first optical fibre <b>17</b>.

The entry of the light into the multimode fibre of the first optical fibre <b>17</b> determines here the number and distribution of the modes stimulated in the multimode fibre. For particularly efficient destruction of the spatial coherence it is advantageous to chose an injection with which the largest possible number of modes are stimulated. This can be implemented in particular by—as shown in <figref idrefs="DRAWINGS">FIGS. 10</figref><i>a</i>) and <b>10</b><i>b</i>)—the focus <b>55</b> of the light beams, i.e. of the third partial beam <b>4</b>, not lying on the facet <b>9</b>, i.e. the entry plane, of the multimode fibre of the first optical fibre <b>17</b> and/or by the light beams of the third partial beam <b>4</b> being injected at an angle into the multimode fibre of the first optical fibre <b>17</b>, the optical axis <b>56</b> of the light beams being tilted in relation to the central axis <b>57</b> of the multimode fibre of the first optical fibre <b>17</b> and enclosing an angle ω, which is preferably between 5° and 40°, with the latter. In this way on the one hand the spatial coherence is suppressed to the maximum, and on the other hand the illumination of the facet <b>9</b> of the multimode fibre is more homogeneous.

Moreover, in <figref idrefs="DRAWINGS">FIGS. 10</figref><i>a</i>) and <b>10</b><i>b</i>) the characteristic diameter d of the multimode fibre used in the first optical fibre <b>17</b> is drawn in.

The described combination of the injection of highly-coherent light <b>14</b> into the first interferometer <b>10</b> in combination with the injection of the light of the third partial beam <b>4</b> subsequently spectrally modulated in the first interferometer <b>10</b> into the first optical fibre <b>17</b> makes it possible to form the optics very simply in the region of the output <b>8</b> of the first interferometer <b>10</b>.

Since with this principle bright coherent light sources, such as e.g. SLEDs, short pulse lasers or supercontinuum lasers, can be used as a light source <b>15</b>, it is possible to achieve considerably higher output densities than with the conventionally used temporally incoherent light sources. The signal/noise ratio of the image information obtained is in this way considerably improved.

Alternatively to the free beam interferometer illustrated and described here, by using this principle the first interferometer <b>10</b> can also be designed totally as a fibre interferometer. The depth scan could then be implemented e.g. instead of by means of the movement of the second reference mirror <b>12</b>, by extending a fibre in one of the two arms of the first interferometer <b>10</b> by means of a so-called fibre stretcher.

8. Image Transfer by Means of Fibre Bundles

As already explained in greater detail, with the present OCT system a depth scan is implemented by means of a macroscopic movement of the reference mirror <b>12</b> in the first interferometer <b>10</b> while the light reflected by the specimen <b>1</b> is forwarded to the two-dimensional detector <b>30</b> via the second interferometer <b>20</b> and the second optical fibre <b>29</b> and collected by the latter.

A fibre bundle made up of a plurality of individual fibres is used as a second optical fibre <b>29</b>. Fibre bundles generally have a high numerical aperture which is technically limited and comes within the range of 0.4 or over. Furthermore, the filling factor of the facet, i.e. the inlet or outlet cross-section, of conventional fibre bundles is relatively small. Both would lead to undesired light losses with the transmission of the light reflected by the specimen <b>1</b> from the second interferometer <b>20</b> to the detector <b>30</b>.

In order to obtain the most compact possible OCT system with small light and information losses when transmitting the light reflected by the specimen <b>1</b> the fibre bundle described in greater detail in the following is used.

<figref idrefs="DRAWINGS">FIG. 11</figref> shows a section <b>50</b> of the facet of the fibre bundle used which—as can be seen from the partial region <b>51</b> illustrated in enlarged form—is made up of a plurality of individual fibres <b>52</b> which have a centre-centre distance d<b>2</b> (so-called fibre pitch).

<figref idrefs="DRAWINGS">FIG. 12</figref> shows a section of the detector <b>30</b> used which comprises a plurality of detector elements <b>80</b> arranged in an area and which have a centre-centre distance d<b>1</b> (so-called pixel pitch). With the present OCT system the fibre pitch d<b>2</b> of the individual fibres <b>52</b> of the fibre bundle is smaller than the pixel pitch d<b>1</b> of the detector elements <b>80</b> of the detector <b>30</b>.

In order to make possible the largest possible field of vision with high spatial resolution, the fibre bundle comprises at least 100,000, preferably approximately 300,000, individual fibres <b>52</b>. The number of detector elements <b>80</b> of the detector <b>30</b> is preferably approximately 328,000 and thus comes within the same order of magnitude as the number of individual fibres <b>52</b>.

As shown in <figref idrefs="DRAWINGS">FIG. 13</figref>, the form of the cross-section of the fibre bundle of the second optical fibre <b>29</b> in the region of the inlet and outlet surface <b>7</b> and <b>6</b> is preferably adapted to the geometry of the detector <b>30</b>, in particular the form of the inlet surface <b>7</b> on the side of the second interferometer <b>20</b> being substantially equal to the form of the outlet surface <b>6</b> on the side of the detector objective <b>31</b> and the detector <b>30</b> (see also <figref idrefs="DRAWINGS">FIG. 1</figref>). The respective form of the inlet and outlet surface <b>7</b> and <b>6</b>, in particular the side length ratio of the latter, is essentially identical here to the preferably rectangular form of the detector <b>30</b>.

In <figref idrefs="DRAWINGS">FIG. 14</figref><i>a</i>) two individual fibres <b>52</b> of the fibre bundle are shown as an example. The individual fibres <b>52</b> have a fibre core <b>65</b> and a fibre cladding <b>66</b>. With the preferably used individual fibres <b>52</b> of the fibre bundle the ratio d<b>3</b>/d<b>4</b> of the thicknesses d<b>3</b> and d<b>4</b> of the respective fibre core <b>65</b> to the fibre cladding <b>66</b> (the so-called core/cladding ratio) is chosen such that the highest possible filling factor is produced with the smallest possible light losses due to light passing out of the fibre <b>52</b> to the side (so-called evanescent waves). The filling factor here is given by the ratio of the whole cross-sectional surface of the individual fibre <b>52</b> to the surface of the fibre core <b>65</b>.

With a wavelength of the light <b>14</b> of for example 1300 nm the fibre bundle used preferably has a fibre pitch d<b>2</b> of 11 μm, a cladding thickness d<b>4</b> of the individual fibres <b>52</b> of 1.7 μm and a core diameter d<b>3</b> of 6.8 μm. The diameter of the individual fibre <b>52</b>, which is produced from the sum of the core diameter d<b>3</b> and twice the cladding thickness d<b>4</b>, is in this case 10.2 μm and is therefore somewhat smaller than the fibre pitch d<b>2</b> because with the production process of the fibre bundle another second cladding (not shown) is produced around each individual fibre <b>52</b>.

In <figref idrefs="DRAWINGS">FIG. 14</figref><i>b</i>) a version of the embodiment of the individual fibres <b>52</b> shown in <figref idrefs="DRAWINGS">FIG. 14</figref><i>a</i>) is illustrated. In this version the individual fibre cores <b>65</b> of the individual fibres <b>52</b> are embedded into a matrix <b>66</b> made of glass or plastic which respectively forms the fibre cladding of each individual fibre core <b>65</b>. With this version two respective adjacent individual fibres <b>52</b> have part of their fibre cladding in common. The distance d<b>4</b> between adjacent fibre cores <b>64</b>, which corresponds to the cladding thickness, can in this way be reduced relative to the individual fibres described above with a respective own fibre cladding, the occurrence of evanescent waves furthermore being efficiently suppressed. The surface ratio of the fibre core surface to the whole fibre surface is in this way particularly large. The quotient of the core diameter d<b>3</b> and the cladding thickness d<b>4</b> here comes within the range between approximately 5 and 8.

The second interferometer <b>20</b> is designed such that for all scan depths a lateral interference pattern is produced the spatial frequency of which is lower than the spatial frequency of the individual fibres <b>52</b> of the fibre bundle, the Nyquist condition in particular having to be fulfilled. This is illustrated in <figref idrefs="DRAWINGS">FIG. 15</figref>. As can be seen in the enlarged section <b>61</b> of the lateral interference pattern <b>60</b>, the length of a period between two consecutive interference minima <b>63</b> (dark rings) of the interference pattern <b>60</b> is greater by a multiple than the centre/centre distance (fibre pitch) of the individual fibres <b>52</b> of the fibre bundle the inlet surface <b>6</b> of which (see also <figref idrefs="DRAWINGS">FIG. 1</figref>) is illustrated here as a section and correspondingly enlarged. Correspondingly, the spatial frequency of the interference pattern <b>60</b> is considerably lower than the spatial frequency of the individual fibres <b>52</b> of the fibre bundle.

With respect to systems known from the prior art wherein the detector is incorporated into the interferometer, a number of advantages are achieved by using the fibre bundle described above which will be described in greater detail in the following.

The pixel pitch d<b>1</b> of InGaAs CMOS detectors, which are sensitive to light with wavelengths within the range of approximately 1300 nm, can not be substantially smaller than 20 μm for technical reasons. The fibre bundle preferably used in the present OCT system has a fibre pitch d<b>2</b> of 10 μm and therefore with the same resolution has a substantially smaller cross-section than the detector. This enables a considerably more compact design of the measuring head in comparison to systems wherein the detector is incorporated into the measuring head.

Moreover, with the aforementioned systems from the prior art, due to the very high sampling rates required of the detector, transmission of data at extremely high speed from the measuring head to the downstream electronics would be required. Moreover, A/D converters would have to be integrated into the measuring head. These disadvantages do not apply to the forwarding described here of the image information obtained from the specimen <b>1</b> by means of the second optical fibre <b>29</b> in the form of a fibre bundle to a detector <b>30</b> separate from the second interferometer <b>20</b>.

Since with the present OCT system no electronics are required, therefore, in order to record and/or process the image in the measuring head, there is no lost heat which could lead to undesired heating of the measuring head.

Since in the second optical fibre <b>29</b> a fibre pitch d<b>2</b> (e.g. 11 μm) is preferably chosen which is smaller than the smallest possible pixel pitch d<b>1</b> (mainly larger than or equal to 20 μm) of the detector <b>30</b>, an enlargement of the image obtained from the specimen <b>1</b> in the measuring head with equal lateral resolution in comparison to systems from the prior art can be reduced, and this makes more simple and smaller optics possible in the second interferometer <b>20</b>.

In order to increase the light yield with the light and image information transmission from the specimen <b>1</b> or from the third reference mirror <b>25</b> to the detector <b>30</b>, adaptation of the numerical apertures of individual components of the present OCT systems is provided, in particular of the apertures of the specimen objective <b>41</b> and of the lenses <b>47</b> in the output arm <b>27</b> and of the apertures of the reference objective <b>46</b> and of the fibre bundle of the second optical fibre <b>29</b>, of the detector objective <b>31</b> and of the detector <b>30</b>. This is described in greater detail in the following by means of <figref idrefs="DRAWINGS">FIGS. 1</figref>, <b>4</b> and <b>16</b>.

<figref idrefs="DRAWINGS">FIG. 16</figref> shows a section of the second optical fibre <b>29</b> made up of a plurality of individual fibres <b>52</b> in the region of the inlet surface <b>7</b>. A convergent light bundle <b>58</b> passing out of the second interferometer <b>20</b> has an aperture angle a and strikes the optical fibre <b>29</b> at an angle of incidence β in relation to the perpendicular of the inlet surface <b>7</b>. The individual fibres <b>52</b> of the second optical fibre <b>27</b> have an aperture angle γ within which they can collect arriving light. The aperture angle γ is given by the numerical aperture of the individual fibres <b>52</b>.

In order to guarantee the highest possible light yield provision is preferably made such that the sum of the aperture angle α of the light bundle <b>58</b> and the angle of incidence β is smaller than the aperture angle γ of the individual fibres <b>52</b> of the fibre bundle <b>29</b>: α+β&lt;γ. In this way it is guaranteed that all of the light of the light bundle <b>58</b> which strikes an individual fibre <b>52</b> passes into the latter and is conveyed to the outlet surface <b>6</b> of the second optical fibre <b>29</b>.

The aperture angle α and the angle of incidence β of the light bundle <b>58</b> required for this are produced by a corresponding embodiment of the specimen and/or reference and/or output objective <b>41</b>, <b>46</b> and <b>47</b>. This is achieved in particular by the two objective combinations of the specimen and output objective <b>41</b>/<b>47</b> or the reference and output objective <b>46</b>/<b>47</b> imaging in enlarged form, i.e. the aperture angle α of the light bundle <b>58</b> in the region of the inlet surface <b>7</b> of the fibre bundle (“image side”) is smaller than the aperture angle (not shown) on the side of the specimen <b>1</b> (“object side”). In this way a large aperture angle can be easily produced on the side of the specimen <b>1</b> by means of which high light collecting efficiency is achieved. Together with the loss-free injection of light into the fibre bundle of the second optical fibre <b>29</b>, in this way overall a very high light yield is guaranteed when collecting the light reflected by the specimen <b>1</b>, and so a high image quality is achieved.

Alternatively, or additionally, in order to increase the light yield, adaptation of the fibre bundle side numerical aperture of the detector objective <b>31</b> to the numerical aperture of the fibre bundle of the second optical fibre <b>29</b> is provided. The aperture angle of the detector objective <b>31</b> is greater here than the aperture angle γ of the individual fibres <b>52</b> of the fibre bundle.

Preferably, the detector objective <b>31</b> is telecentric on the side of the fibre bundle. In this way the radiation characteristics of the fibre bundle can easily be allowed for. The field angle on the output surface <b>6</b> is equal to zero for every position on the output surface <b>6</b>.

As the angle of incidence of the light beams onto the detector <b>30</b> increases, the light output collected by the detector <b>30</b> becomes smaller. In order to guarantee the highest possible light yield provision is therefore made such that the angle of incidence of the light beams onto the detector <b>30</b> is kept as small as possible. This is preferably achieved by enlarged imaging of the fibre bundle of the second optical fibre <b>29</b> onto the detector <b>30</b> and a telecentric design of the detector objective <b>31</b> on the side of the detector <b>30</b>.

A further advantage when using the described fibre bundle for image transmission is that the overall enlargement M of the system can be split into two steps, namely into a first enlargement M<b>1</b> in the measuring head, i.e. in the second interferometer <b>20</b>, and a second enlargement M<b>2</b> in the detector objective <b>31</b>. In this way the first enlargement M<b>1</b> of the objectives <b>41</b>, <b>46</b> and <b>47</b> in the measuring head can be smaller than the overall enlargement M required for the nominal resolution of the OCT system. The following example is intended to illustrate this: With a pixel pitch of 20 μm, a fibre pitch of 10 μm and a nominal resolution of 2.5 μm, by means of the fibre bundle of the second optical fibre <b>29</b> formed as described above, an enlargement M<b>1</b>=4 is produced in the measuring head and an enlargement M<b>2</b>=2 in the detector objective <b>31</b> so as to obtain an overall enlargement M=M<b>1</b>×M<b>2</b>=8. Without an image transmission by means of the described fibre bundle an enlargement equal to the overall enlargement M=8 would, however, have to be produced in the measuring head.

Therefore, the advantage of using the fibre bundle described above is that the overall enlargement M does not only have to be provided by the objectives of the second interferometer <b>20</b>, and so the specimen and/or reference and/or output objectives <b>41</b>, <b>46</b> and <b>47</b> of the measuring head can be simpler and more space-saving in design, by means of which the measuring head can be substantially more compact in design overall.

As in the example of a second interferometer <b>20</b> shown in <figref idrefs="DRAWINGS">FIG. 4</figref>, in this way the average diameter D<b>1</b> of the specimen objective <b>41</b> and of the lenses <b>47</b> of the output objective of the second interferometer <b>20</b> can preferably be chosen to be smaller than the diameter D<b>2</b> of the second optical fibre <b>29</b> in the region of the inlet surface <b>7</b>: D<b>1</b>&lt;D<b>2</b>.

9. Operating Modes of the OCT System

The OCT system described above can be operated in three different operating modes. The operating modes are two real time modes, wherein OCT images of a specimen are produced at a high rate of approximately 5 to 10 images per second, and a static operating mode.

In the first operating mode, the real time mode <b>1</b>, two-dimensional depth sections of the specimen <b>1</b> (so-called slices) are produced in real time. This is implemented in that as a detector <b>30</b> a CMOS camera is used which allows a so-called Window of Interest (WOI) to be set with which only one partial surface of the detector <b>30</b> is sensitive to light and converts the latter into corresponding detector signals. The reduction of the sensitive camera surface is associated with a considerable increase in the camera speed; with this setting more camera images can be produced per second than in the complete image mode.

In the real time mode <b>1</b> a WOI is preferably chosen which in one direction corresponds to the whole camera length and width (e.g. 640 pixels) and in the other direction has the minimum possible number of pixels (e.g. 4 pixels) as given by the type of respective camera. In this way the speed of the camera is increased to such an extent that OCT images can be recorded in real time.

This is preferably achieved in combination with the modulation of the intensity of the light <b>14</b> or <b>4</b> injected into the first interferometer <b>10</b> and emitted by the first interferometer <b>10</b> or the modulation of the sensitivity of the detector system <b>30</b>, <b>31</b> (see sections 3 and 4 above).

<figref idrefs="DRAWINGS">FIG. 17</figref> shows a detector surface F<b>1</b> which is made up of a first number N<b>1</b> of detector elements <b>80</b> and has a length c<b>1</b> and a width b<b>1</b>. With the aforementioned setting of a WOI, light is only collected by the detector elements <b>80</b> located in a partial surface F<b>2</b> of the detector surface F<b>1</b> and is converted into corresponding detector signals. The second number N<b>2</b> of detector elements <b>80</b> of the partial surface F<b>2</b> is smaller than the first number N<b>1</b> of detector elements <b>80</b> of the whole detector surface F<b>1</b>. The lengths c<b>1</b> and c<b>2</b> of the detector surface F<b>1</b> and partial surface F<b>2</b> are of equal size, whereas the widths b<b>1</b> and b<b>2</b> of the detector surface F<b>1</b> and partial surface F<b>2</b> are different.

In the example shown the partial surface F<b>2</b> is only four pixels wide, whereas the detector surface F<b>1</b> is 512 pixels wide. The sensitive surface of the detector surface F<b>1</b> is therefore reduced by a factor of 128, and this considerably reduces the period of time required for the recording of interference patterns and conversion of the latter into corresponding detector signals.

As shown in <figref idrefs="DRAWINGS">FIG. 18</figref>, in this example, instead of a complete three-dimensional tomogram, only four (corresponding to the four rows of pixels of the partial surface F<b>2</b>) two-dimensional depth sections <b>67</b> are obtained from the observed spatial element <b>33</b> of the specimen <b>1</b>.

In the second operating mode, the real time mode <b>2</b>,—as shown in FIG. <b>19</b>—two-dimensional tomograms <b>68</b> are produced from a specific depth T of the observed spatial element <b>33</b> of the specimen <b>1</b>, it being possible to choose any depth T. Here the whole detector surface F<b>1</b> of the detector <b>30</b> is used for collecting the light reflected by the specimen <b>1</b> and the conversion of the latter into corresponding detector signals, only a maximum of five camera images, however, respectively being used to calculate a tomogram <b>68</b>. For this purpose the first reference mirror <b>11</b> is periodically moved within the first interferometer <b>10</b> with an amplitude of approximately 1 μm, whereas up to five camera images are recorded which are then allocated to one OCT image. In this way tomograms <b>68</b> with a high repetition rate can be produced.

By means of a macroscopic movement of the second reference mirror <b>12</b>, optionally in combination with the focus tracking (see section 1 and 2 above), any depth T from which the tomogram <b>68</b> is obtained can be chosen.

In the third operating mode, the static mode, a complete three-dimensional set of data is recorded with the aid of the macroscopic movement of the second reference mirror <b>12</b> in combination with the focus tracking. Details with regard to this can be taken in particular from sections 1 and 2.

By means of the different operating modes the OCT system can fulfil a whole range of different requirements. The functionalities when examining specimens, for example when locating relevant points in the specimen, are thus considerably extended.

10. Further Inventive Aspects of the System and Method for OCT

The system and method for OCT described in greater detail above has individual features or combinations of features by means of which the system and method are made simpler and more compact in design and quicker and more reliable when handling and image recording without all of the features listed in the preamble and/or characterising part of the independent claims being required. These features and combinations of features are also considered to be the invention.

The invention is considered in particular to be a system for optical coherence tomography with <ul><li id="ul0003-0001" num="0000"><ul><li id="ul0004-0001" num="0226">at least one interferometer for emitting light with which a specimen is irradiated, and</li><li id="ul0004-0002" num="0227">a detector for collecting light which is reflected by the specimen, <br /> the system being characterised by one or a number of features which have been described in greater detail above, in particular in sections 1 to 9 and/or in connection with <figref idrefs="DRAWINGS">FIGS. 1 to 19</figref>. </li></ul></li></ul>

The method corresponding to this system is also considered to be the invention.

Irradiation of the specimen with light emitted by the interferometer takes place either indirectly, i.e. by means of a further interferometer which is located between the interferometer and the specimen, or directly, i.e. without a further interferometer located between the interferometer and the specimen.

The collection by the detector of the light reflected by the specimen takes place either indirectly, i.e. by means of a further interferometer which is located between the specimen and the detector, or directly, i.e. without a further interferometer located between the detector and the specimen.